384 Mr. C. Meldrum on [Mar. 9, 



the maxima to the minima epochs, and the ratios of the former to the 

 latter intervals are nearly the same for all the three cycles. 



(20.) These results will be somewhat modified by the fourth and fifth 

 terms of the equations and by additional observations ; but at present 

 they may be considered fair approximations. The mean excess of 

 rainfall (6*19 inches) is almost the mean of the 5*19 and 7*20 inches 

 previously obtained, and the 16*73 inches of river-fluctuation are nearly 

 the same as the 16*20 inches found by taking the means of the depths in 

 the adopted maxima and minima years. The durations of the periods, 

 and the intervals from minimum to maximum and from maximum to 

 minimum, given by the equations are also not very different from those 

 given directly by the observations. 



(21.) Although the mean depths of the rivers have apparently been 

 decreasing, yet the range of variation has altered little. It would thus 

 appear that there is some general cause which operates irrespectively of 

 local causes, and that its effects usually override all other effects. This 

 is shown by the differences between the greatest and least depths for 

 consecutive periods of ten to thirteen years. 



The more numerous the observations the more evident does a con- 

 nexion between sun-spots and rainfall become. At first (1872) the 

 rainfall at only three stations (Mauritius, Adelaide, and Brisbane) was 

 examined. To these Mr. J. Norman Lockyer, in a paper " On the 

 Meteorology of the Future " (' Nature,' vol. vii. p. 142), added the rain- 

 falls of the Cape and of Madras, strongly expressing in favour of the 

 hypothesis (even at that early stage) an opinion which, though not then 

 accepted by some meteorologists, all subsequent investigations, including 

 those of Mr. Blanford, Prof. Broklesby, Mr. Dawson, Mr. Ellery, Dr. 

 Jelinek, Mr. E. H. Scott, and Mr. Symons, have amply justified*. The 

 number of stations has, in short, gradually risen from three to 144, and 

 the results of each successive investigation have been an improvement on 

 their predecessors. 



Not only has the number of stations been increased; the periods 

 of observation have been greatly extended. At first we had only ten to 

 forty-five years' observations, comprising at the utmost one to four sun- 

 spot periods. We have now rainfall observations for seven and river- 

 observations for thirteen sun-spot periods ; and the fact that the law 

 holds good with the earlier as well as with the later observations is signi- 

 ficant. The variations in the level of the Elbe, from 1728 to 1777, 

 followed the sun-spot variations as regularly as those of the rainfall and 

 rivers have done in later times. 



The amount of rainfall variation is greatest within the tropics, and 

 seems to decrease as the latitude increases. This is a point, how- 

 ever, which requires further examination by considering the observations 

 * I have not seen Prof. Tyndall andMr.JIennessy's contributions. 



